During the first year of this project, efforts were made to digitize and process left ventricular margins using automated computer techniques. Because there have been substantial advances in the area of automatic processing of ventriculograms, our attention has been focused on automated processing of coronary arteriograms. Digitized images from coronary arteriographic cine frames will be computer processed using methods of subtraction, contrast enhancement, and edge finding. Using a series of specialized algorithms, including a skeletonizing transformation, an identification of the coronary vascular tree in a single plane will be identified. A set of logical rules will be developed to define branch points as distinct from intersections. Utilizing this single planer identification of the coronary tree, multiple projections will be analyzed to provide a three-dimensional projection of the coronary tree. In addition, with the identification of the vascular distribution made, it becomes possible to turn to the original images and extract more detailed information concerning vessel diameters and degree of stenosis.